Fri 14 Oct 2016 10:46:48 PM UTC, comment #5:
I renamed some of the obvious Octave-only properties to have underscores, or more closely match Matlab's own hidden property names.
See http://hg.savannah.gnu.org/hgweb/octave/rev/4191f8b5070c.
See http://hg.savannah.gnu.org/hgweb/octave/rev/76f1164d2135.
This still leaves the more delicate work of renaming the hidden "alim", "clim", "xlim", "ylim", and "zlim" properties for the image, patch, surface, and hggroup objects.
|
Tue 11 Oct 2016 08:16:09 PM UTC, comment #4:
We have to be careful. Many "hidden" properties are also hidden in Matlab, but they are so well known that people expect to find them under the same names. Examples include 'xliminclude', 'looseinset', 'x_viewtransform'. Properties which strictly belong to octave, like 'fltk_label', should be renamed to have underscores.
Can someone run the following in Matlab to see if these properties are known in Matlab?
|
Wed 28 Sep 2016 03:57:54 PM UTC, comment #3:
Yes, you are right that it is easy to mistakenly get an image "clim" instead of its parent axes'. Renaming hidden properties would avoid confusion.
I'll retitle this report and make it a feature request.
|
Wed 28 Sep 2016 03:25:17 PM UTC, comment #2:
OK, this hidden property rings a bell. Probably using something like _clim_, _xlim_, etc. for the internal property name would have been better, as it is very easy for someone to stumble upon the object clim, xlim, etc.
Looking at graphics.cc, I'm guessing that these x/y/clim object properties are
min_val, max_val, min_pos, max_neg
So a max_neg of -Inf means there are no negative numbers in the data.
|
Wed 28 Sep 2016 09:35:42 AM UTC, comment #1:
"clim" image property is an internal, hidden, non-documented, non matlab compatible property, see:
"clim" is used internally to compute the image parent axes limits, but is not meant to be used outside the graphics engine. So I think this report can be closed.
|
Wed 28 Sep 2016 08:34:58 AM UTC, original submission:
With the following set of commands there is a clim that has four elements, but I think only two elements are valid (toolkit doesn't matter):
The 2 and 255 seem correct. I don't know what 2 -Inf might represent. Out of curiosity:
|